lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jeremy Ashcraft <>
Subject Re: solr blocking and client timeout issue
Date Sat, 18 Jul 2015 06:42:55 GMT
I turned on GC logging and verified that its definitely being caused by 
a GC pause.  I tried the tuning option from the article and get this 

  OpenJDK 64-Bit Server VM warning: Failed to reserve shared memory 
(errno = 1).

any recommendations on how to get rid of that warning, and should I be 
worried about it?

On 7/17/2015 4:14 PM, Shawn Heisey wrote:
> With a 4GB or 5GB heap, I would *definitely* do some GC tuning. I 
> would also strongly recommend upgrading your Java version to at least 
> the latest Java 7. This is what I use for getting the latest Java 7 
> version onto CentOS 6. It is an excellent package that puts relevant 
> versions of all the important commandline tools into your path: 
> There aren't really 
> any good Java 8 packages for CentOS 6, so you may be stuck with Java 
> 7, but if you *can* go to the latest Java 8, there are rumored to be 
> some big performance gains due to better memory management. For GC 
> tuning, I have a page in the Solr wiki with tuning options that have 
> been working very well for me: 
> If you want to be 
> conservative, skip the G1 tuning and go for the CMS tuning options. We 
> also have a page dedicated to performance problems with Solr, which 
> can, as I've described, be related to Java heap issues. One of the 
> things on that page is a section dedicated to reducing your heap 
> requirements, which you can find in the Table Of Contents near the 
> top: If you have 
> some time, reading that entire page is a good idea. Thanks, Shawn 

View raw message